Community
Participate
Working Groups
I have the following advice Object around() : callingThirdParty() && ! call(java.lang.Throwable+.*(..)) // cause trouble && ! call(java.util.logging..*.*(..)) // cause trouble && ! call(String toString()) && ! call(void log(..)) { ... } I am not sure whether the syntax is allowed, but it should not throw exception is a syntax error is detected. If I removed both troublsome lines, the compilation would be successful. Either of the lines will cause the following exception. Checked bug rebort database and found one bug from the same line of code, but the fix comment seems totally different. compiling D:\WebSphere\AppServer60\aspects\trace\TraceMyClasses.java ABORT Exception thrown from AspectJ 1.2.1 This might be logged as a bug already -- find current bugs at http://bugs.eclipse.org/bugs/buglist.cgi?product=AspectJ&component=Compiler Bugs for exceptions thrown have titles File:line from the top stack, e.g., "SomeFile.java:243" If you don't find the exception below in a bug, please add a new bug at http://bugs.eclipse.org/bugs/enter_bug.cgi?product=AspectJ To make the bug a priority, please include a test program that can reproduce this exception. null java.lang.NullPointerException at org.aspectj.weaver.patterns.PerFromSuper.concretize(PerFromSuper.ja :59) at org.aspectj.weaver.CrosscuttingMembers.setPerClause(CrosscuttingMem rs.java:207) at org.aspectj.weaver.ResolvedTypeX.collectCrosscuttingMembers(Resolve ypeX.java:329) at org.aspectj.weaver.CrosscuttingMembersSet.addOrReplaceAspect(Crossc tingMembersSet.java:57) at org.aspectj.weaver.bcel.BcelWeaver.prepareForWeave(BcelWeaver.java: 3) at org.aspectj.ajdt.internal.compiler.AjCompilerAdapter.weave(AjCompil Adapter.java:226) at org.aspectj.ajdt.internal.compiler.AjCompilerAdapter.afterCompiling jCompilerAdapter.java:118) at org.aspectj.org.eclipse.jdt.internal.compiler.Compiler.compile(Comp er.java:383) at org.aspectj.ajdt.internal.core.builder.AjBuildManager.performCompil ion(AjBuildManager.java:680) at org.aspectj.ajdt.internal.core.builder.AjBuildManager.doBuild(AjBui Manager.java:168) at org.aspectj.ajdt.internal.core.builder.AjBuildManager.batchBuild(Aj ildManager.java:102) at org.aspectj.ajdt.ajc.AjdtCommand.doCommand(AjdtCommand.java:109) at org.aspectj.ajdt.ajc.AjdtCommand.runCommand(AjdtCommand.java:60) at org.aspectj.tools.ajc.Main.run(Main.java:291) at org.aspectj.tools.ajc.Main.runMain(Main.java:227) at org.aspectj.tools.ajc.Main.main(Main.java:80) Dumping to ajcore.20050311.233710.655.txt 1 fail|abort
This is a definite bug, you should never see an NPE coming out of the compiler. If you change the advice to: Object around() : callingThirdParty() && ! call(* java.lang.Throwable+.*(..)) // cause trouble && ! call(* java.util.logging..*.*(..)) // cause trouble && ! call(String toString()) && ! call(void log(..)) { ... } does your program now compile correctly or do you see the same / similar error? (There is a syntax error in your pointcut expression - you need to specify a return type for the first two calls, but that should give you a compilation error, not an NPE...)
scheduling for investigation in aj5m3
Hmmm. I created this test program: aspect C { Object around() : callingThirdParty() && ! call(java.lang.Throwable+.*(..)) // cause trouble && ! call(java.util.logging..*.*(..)) // cause trouble && ! call(String toString()) && ! call(void log(..)) { } public void log() { } } and under 1.2/1.2.1/1.5.0M2 I always got the kind of error you would normally expect for getting the call sig wrong: K:\ws\aspectj_ws\tests\bugs150\PR87297.aj:4 [error] Syntax error on token ".", "expected name pattern" expected && ! call(java.lang.Throwable+.*(..)) // cause trouble Given the location of the NPE, I think there is also another part of the aspect involved in the failure - did you have a per clause defined for it? The failure appears to happen when there is a problem with the perclause - I can imagine a situation where the syntax error for your pointcut occurs and it leaves the perclause in a funny state which then fails NPE - but I can't imagine what kind of perclause could be the problem? I don't want to put some guards in for the NPE without knowing the failing per clause that causes the problem.
still waiting to hear on the perclause...
closing this bug out for now as per andy's last two comments. Please reopen with reproduceable test case if it is still an issue for you.... thanks, A.
